Eastern Analytical Symposium 2007
November 12–14, 2007
Garden State Convention Center
Somerset, NJ, USA
Conference Details
Booth # 321
| |
Poster Schedule
|
| Title: |
A System for Efficient Chiral Chromatographic Method Development and Knowledge-base Accumulation |
| Authors: |
Mike McBrien & David Snyderman (ACD/Labs) and Adam Beard & Xiaoyi Gong (Merck) |
| Date: |
Wednesday, November 14, 2007 @ 12:00–2:00 PM |
| Session Title: |
HPLC and Chiral Chromatography |
| Location: |
Exhibit Hall |
| Presenter: |
Graham McGibbon |
| Abstract: |
One of the key chromatographic challenges in both drug discovery and development is the purification of enantiomeric compounds. Development of purification methods can entail a great deal of effort due to the large number of separation systems available, and the high-throughput requirements of the modern drug world. Most laboratories have attempted to address this challenge through the application of more, and faster, apparati for screening chiral selectivities. Once a reasonable selectivity has been exhibited, some simple optimization principles may be applied, and the sample is purified. This experiment-intensive approach facilitates fast method development, but requires considerable resources in terms of instrumentation and manual data processing and review time. In addition, it is common to have situations where the results of chiral selectivities screens are not retained such that other members of the organization are forced to perform duplicate experimentation when prior results are not available.
This paper will present a system for fast chiral selectivity data review. The system involves automated linking of chromatograms, chromatographic methods, and chemical structures. Chiral selectivities are automatically calculated after data has been collected by the instrument. After the chromatographer reviews the results, a method is targeted for optimizing and/or scale-up. Finally, the reviewed selectivities are automatically moved to the central database. Thus all chromatographic knowledge collected by the laboratory becomes part of a permanent organizational knowledgebase. Additionally, the linking of chemical structures with applicable methods enables the reduction of the number of experiments that are required for future compounds by short-listing screens based on aspects of structural-similarity.
|
Talk Schedule
|
| Title: |
Isotope Information in LC-MS Data Sets Aids Automated Software Algorithms to Effectively Recognize, Extract and Label Peaks of Sample Components |
| Authors: |
Graham A. McGibbon, Margaret Antler, Mark A. Bayliss, Vitaly Lashin (ACD/Labs) |
| Date: |
Tuesday, November 13, 2007 @ 10:40 AM |
| Location: |
Ballroom Center |
| Presenter: |
Graham McGibbon |
| Session Title: |
Strategies and Applications of Liquid Chromatography—Mass Spectrometry |
| Abstract: |
A common concern for qualitative LC-MS analyses of samples containing multiple compounds, related or not and varying in quantity, is to effectively identify all the significant components. Considerable alleviation is available via software automation that helps overcome the challenge of finding all the peaks in LC-MS data sets, including identifying differences between multiple related samples. Post-acquisition algorithms offer the potential for effective data management, processing, and component annotation if they can automatically perform a thorough analysis of each data set: extracting and determining those m/z values that give rise to chromatographic peaks. Recognizing the presence of the isotopes of ionized molecules from the sample and including these peaks bolsters assignment of sample components. Nevertheless, for automated detection of components with peaks of small signal-to-noise ratio, which may represent the most interesting (bio)chemical transformations, comprehensiveness and discernment are both vital. We will discuss aspects of a unique algorithm for componentization of LC/MS data that has been developed to facilitate automated analysis of extracted ion chromatogram peaks including consideration of isotope distribution information in the mass spectra to ensure that all features relevant to a chemical species are identified. Application of the software enables concurrent reduction of human time and effort expended in reviewing and evaluating data and results. |
| |
| Title: |
Interpreted IR and Raman Databases |
| Authors: |
Michael Boruta & Michel Hachey (ACD/Labs) and Gene Hall (Rutgers University) |
| Date: |
Wednesday, November 14, 2007 @ 9:20 AM |
| Location: |
Princeton Room |
| Presenter: |
Michael Boruta |
| Session Title: |
Novel Topics in Spectroscopy |
| Abstract: |
When searching IR or Raman spectra against spectral databases, not finding an exact match it is more common than we would like. This is less frequent than one would expect with over 10,000,000 compounds in the world and some 100s of thousands of reference spectra. However, just because an exact match is not found does not mean that the search did not provide us with helpful information. Often there is useful information about the classification of the unknown and some assistance in the interpretation; the difficulty is in extracting that information. Having a database that includes spectral structure correlations can make that task easier. This talk will report on an ongoing project to develop interpreted IR and Raman databases to assist in the identification, interpretation, and classification of results form spectra searching. |
| |
| Title: |
Robustness of Variable Selection Based on Pure Variables |
| Authors: |
Michel Hachey, Michael Boruta (ACD/Labs) |
| Date: |
Wednesday, November 14, 2007 @ 10:40 AM |
| Location: |
Tewksbury Room |
| Presenter: |
Michel Hachey |
| Session Title: |
Chemometrics—Finding the Right Method for Your Application |
| Abstract: |
Since many wavelength selection methods can produce wildly different wavelength suggestions providing small changes in the data set, it is interesting to examine how robust the recently proposed pure variable method for selecting wavelength is. In this method, the variable selection method is guided by the purity function from the SIMPLe-to-use Interactive Self-modeling Mixture Analysis (SIMPLISMA) algorithm to help determine the most significant and influential regions in a multivariate calibration. A series of publicly available data sets will be used to study the robustness and help differentiate this method from alternate wavelength selection strategies that are more statistically focused. |
| |
| Title: |
Managing Data in LC-MS/DAD Forced Degradation Studies |
| Authors: |
Mike McBrien, Margaret Antler, Andrey Vazhentsev, Vadim Tashlitsky, and Alexey Galin (ACD/Labs) |
| Date: |
Thursday, November 15, 2007 @ 11:20 AM |
| Location: |
Ballroom East |
| Presenter: |
Graham McGibbon |
| Session Title: |
Current Strategies of HPLC Method Development |
| Abstract: |
Today's chromatographers have an unprecedented capacity for generation and collection of data associated with unknown samples. The combination of immense hard drives, multiplexed detection, and robust instrumentation provides the analyst with a tremendous amount of data to manage.
One of the largest challenges facing modern chromatographers is the problem of development of methods for detection and quantitation of impurities in samples associated with drug stability studies. A large number of potential impurities can be generated through various reactive conditions, including exposure to UV light, acid, base, peroxide, etc. Chromatographers then design chromatographic conditions for quantitation of the impurities, injecting multiple related samples for each set of conditions individually. The components are used collectively in method development. This eases the burden of detection of low-level species, and of chromatographic peak tracking across sets of conditions. Thus multiple, relatively simple samples represent one more complex composite sample for purposes of generation of a separation method.
The result of combining this composite sample approach with a systematic strategy of investigating its chromatographic response can be a very rigorous determination of optimal chromatographic conditions. However, the amount of data that is generated, and the time required to analyze it, can be onerous. Chromatographic peak tracking and cross-sample component resolution, and the necessity of transcribing peak tables into modeling software, can require more time than the data collection itself. This paper will describe the application of new techniques for management of method development data, both for composite, and "monotone" samples, and the utilization of this data for purposes of rigorous method development.
|
For complete event information, visit EAS.
|
